BA/MA/IDP: Computer-aided survival, grading prediction and segmentation of soft tissue sarcomas in MRI.
Advisor: Prof. Bjoern Menze
Supervision by: Fernando Navarro
This project will investigate three-main task in soft-tissue sarcomas. The first is the prediction of disease progression or survival given different time point scans of the patient. The second consist of grading the aggressiveness of the sarcomas in a three-class classification task. The third task is related to the medical image segmentation to automate the treatment planning.
Figure 1: Project description illustration. The system must first segment the soft-tissue sarcoma, followed by a grading prediction of the aggressiveness of the tumor and the disease survival or progression.
Background and Motivation
Soft-tissue sarcomas are an aggressive type of tumor that develops in patient’s soft-tissue without noticeable side effects until it is detected in advance stages. Once it is detected, patients receive MRI studies for treatment planning and estimation of patient's outcome. To assess the aggressiveness of a tumor, invasive biopsies are additionally necessary. Using current prognostic models and therapies about half of all patients still suffer from tumor spread throughout the body. With a novel computer-aided system to automatically analyze and segment soft-tissue sarcomas based solely on the MRI scan, which is a non-invasive imaging modality, may help to advance patient care.
Disease progression and survival is a well-known topic and widely investigated in brain tumor . Most research focuses on analyzing radiomic features and their relationship to MRI features as described in [2,3]. Nevertheless, extraction of radiomic features is an expensive and time-consuming task leading to an unsuitable solution for real medical applications. As an alternative MRI features are a potential surrogate solution, where the final goal is to be able to analyze soft-tissue sarcomas given an MRI of the patient. To the best of our knowledge, this is the first work in this direction and the one we would like to explore in this research project.
The aim of this project is to develop and implement deep neural networks in an end-to-end strategy to perform three-task in analyzing soft-tissue sarcomas. Our clinical partners in the radiotherapy department at klinikum Rechts der Isar and Seattle University Hospital have collected and annotated large dataset on the study of soft-tissue sarcomas. The dataset contains over 200 patients with continuous outcome data, grading of aggressiveness and segmentation of the soft-tissue sarcomas during the period of two years.
The student task would be to perform the following tasks:
- Develop and implement deep neural networks to automatically predict the survival or disease progression of soft-tissue sarcomas given an input MRI scan of the patient.
- Develop and implement classification architectures to grade the aggressiveness of soft-tissue sarcomas based on visual features from MRI.
- Automatic segmentation of soft-tissue sarcomas in MRI scans for radiation treatment planning.
- Experimental evaluation from the different developed systems.
Depending on the desired thesis type and student's preferences, the tasks will be adapted to suit the scope of the project.
- Knowledge of image processing
- Experience with deep learning
- Good python programming skills
- Experience with deep learning frameworks like TensorFlow?
 Nie, D., Zhang, H., Adeli, E., Liu, L. and Shen, D., 2016, October. 3D deep learning for multi-modal imaging-guided survival time prediction of brain tumor patients. In International Conference on Medical Image Computing and Computer-Assisted Intervention (pp. 212-220). Springer, Cham.
 Vallières, M., Freeman, C.R., Skamene, S.R. and El Naqa, I., 2015. A radiomics model from joint FDG-PET and MRI texture features for the prediction of lung metastases in soft-tissue sarcomas of the extremities. Physics in Medicine & Biology, 60(14), p.5471.
 Corino, V.D., Montin, E., Messina, A., Casali, P.G., Gronchi, A., Marchianò, A. and Mainardi, L.T., 2018. Radiomic analysis of soft tissues sarcomas can distinguish intermediate from high‐grade lesions. Journal of Magnetic Resonance Imaging, 47(3), pp.829-840.